What is Data-Driven PR, Part 4: Hypothesizing

At SHIFT, our approach is to apply equal parts art and science to build integrated programs that help brands connect with the people that matter most. But what does the ‘science’ part of communications entail? What does it look like in action?

First and foremost, it means to be data-driven in our planning and execution; to make informed decisions based on data and research. In this series, we examine how to become a more data-driven communications professional.


In the previous step, we transformed our open-ended questions from Part 2 into variables and data sources. In this step, we will now construct the foundation of true data-driven PR: the hypothesis.

Isn’t data the basis of data-driven PR and communications? Yes and no. What we call data-driven today was once called evidence-based decision-making. Data by itself isn’t necessarily helpful; data and insight that fuels decision-making is. To make the transition from raw data to decision-making fodder requires us to use data to prove or disprove something.

In the scientific method, when we set out to prove or disprove something, we are constructing a hypothesis.

What is a hypothesis?

In data-driven PR, a hypothesis is a statement we intend to prove true or false through the use of the scientific method. Let’s return to our question from parts 2 and 3:

What do customers like or dislike about the taste of espresso?

How would we transform this open-ended question into a hypothesis? We would turn it from a question to a statement we will declare true or false.

What do people associate with the taste of espresso? In a short, qualitative look at conversation about espresso, we find:

espresso word cloud.png

We find flavor profiles of strong, bitter, and burnt as most commonly associated with espresso. Thus, we could frame our hypothesis using any of those words:

  • Our customers like the burnt taste of espresso.
  • Our customers like the strong taste of espresso.
  • Our customers like the bitter taste of espresso.
  • Our customers dislike the burnt taste of espresso.
  • Our customers dislike the strong taste of espresso.
  • Our customers dislike the bitter taste of espresso.

Any one of those statements can be proven true or false, and is the foundation for our data-driven investigation.

What makes a bad hypothesis?

The construction of the hypothesis is the foundation of a data-driven investigation, which leads to data-driven PR done well. Thus, it’s important to understand ways in which a hypothesis can go bad.

Compound hypotheses

One of the most common ways a hypothesis goes bad is with compounding. Using the example above, this would be a bad hypothesis:

  • Our customers dislike the burnt, bitter taste of espresso.

The hypothesis above attempts to test two dimensions at the same time; from a logic perspective, either burnt or bitter could be true, or both could be true. There’s no way to isolate either one in the hypothesis. We should, as much as possible, test one thing at a time.

Malformed hypotheses

The other common way a hypothesis goes bad is if it’s formed incorrectly, such as leaving it a question, creating a statement that cannot be definitively proven true or false, or introducing new variables after the design of a question.

  • Maybe our customers enjoy the taste of burned coffee beans. (cannot be proven true or false)
  • Do our customers enjoy the taste of bitter coffee? (not a statement)
  • Our customers would like the taste of espresso with sugar. (adding in new variables not previously defined)

Any of the errors above would create, at best, inaccurate research, and at worst, a complete waste of time.

Next: testing

Now that we’ve established a hypothesis, our next step in the data-driven PR process is to test the hypothesis. We’ll examine ways to test and gather the data in order to prove our hypothesis true or false.

Keep in Touch

Want fresh perspective on communications trends & strategy? Sign up for the SHIFT/ahead newsletter.

Ready to shift ahead?

Let's talk